Development of a Magnetic Nanoparticles-Based Screen-Printed Electrodes (MNPs-SPEs) Biosensor for the Quantification of Ochratoxin A in Cereal and Feed Samples

A rapid and sensitive electrochemical biosensor based on magnetic nanoparticles and screen-printed electrodes (MNPs-SPEs sensor) was developed for the detection of ochratoxin A (OTA) in cereal and feed samples. Different types of magnetic nanoparticles-based ELISA (MNPs-ELISA) were optimized, and the signal detection, as well as sensitivity, was enhanced by the combined use of screen-printed electrodes (SPEs). Under the optimized conditions, the calibration curve of the MNPs-SPEs sensor was y = 0.3372x + 0.8324 (R2 = 0.9805). The linear range of detection and the detection limit were 0.01–0.82 ng/mL and 0.007 ng/mL, respectively. In addition, 50% inhibition (IC50) was detectable at 0.10 ng/mL. The limit of detection (LOD) of this MNPs-SPEs sensor in cereal and feed samples was 0.28 μg/kg. The recovery rates in spiked samples were between 78.7% and 113.5%, and the relative standard deviations (RSDs) were 3.6–9.8%, with the coefficient of variation lower than 15%. Parallel analysis of commercial samples (corn, wheat, and feedstuff) showed a good correlation between MNPs-SPEs sensor and liquid chromatography tandem mass spectrometry (LC/MS-MS). This new method provides a rapid, highly sensitive, and less time-consuming method to determine levels of ochratoxin A in cereal and feedstuff samples.


Introduction
Ochratoxins are secondary metabolites with toxic properties, produced by some species of fungi including Penicillium and Aspergillus and could be easily found in moldy or fermented agricultural products. The most ubiquitous and toxic of the ochratoxins is ochratoxin A (OTA). OTA can be present

Optimizations and Comparisons of Three Variants of MNPs-ELISA
The dilutions of MNPs-anti OTA, MNPs-BSA-OTA, concentration of OTA-BSA-HRP, anti OTA-HRP, and OTA-HRP of the three types of MNPs-ELISA were optimized, and the results are shown in Table 1. The time for the competition reaction of each MNPs-ELISA was 45 min. The limit of detection (LOD), 50% inhibitory concentration (IC50), and detection ranges of the three different types of MNPs-ELISAs were calculated using Microsoft Excel (version 2016) ( Table 2) under the optimized conditions. The LOD is defined as the average signal corresponding to three standard deviations from the signals of mycotoxin-free samples [29], whereas the concentration of target mycotoxin causing 20-80% inhibition is the detection range [30]. The sensitivities, as well as the working ranges, were

Optimizations and Comparisons of Three Variants of MNPs-ELISA
The dilutions of MNPs-anti OTA, MNPs-BSA-OTA, concentration of OTA-BSA-HRP, anti OTA-HRP, and OTA-HRP of the three types of MNPs-ELISA were optimized, and the results are shown in Table 1. The time for the competition reaction of each MNPs-ELISA was 45 min. The limit of detection (LOD), 50% inhibitory concentration (IC50), and detection ranges of the three different types of MNPs-ELISAs were calculated using Microsoft Excel (version 2016) ( Table 2) under the optimized conditions. The LOD is defined as the average signal corresponding to three standard deviations from the signals of mycotoxin-free samples [29], whereas the concentration of target mycotoxin causing 20-80% inhibition is the detection range [30]. The sensitivities, as well as the working ranges, were improved when using the immune-magnetic beads methods ( Figure 1B) compared with the antigen-coated magnetic nanoparticles ( Figure 1A) and improved even further when using the OTA-HRP ( Figure 1C). This result is consistent with the previous study, which showed that for the detection of small-molecule analytes, the conjugation of analytes with HRP can be used for signal amplification [31]. The three calibration curves are presented in Figure 2. improved when using the immune-magnetic beads methods ( Figure 1B) compared with the antigencoated magnetic nanoparticles ( Figure 1A) and improved even further when using the OTA-HRP ( Figure 1C). This result is consistent with the previous study, which showed that for the detection of small-molecule analytes, the conjugation of analytes with HRP can be used for signal amplification [31]. The three calibration curves are presented in Figure 2.

Specificity Study
The MNPs-ELISA based on MNPs-anti OTA and OTA-HRP had low cross-reactivity with the OTA analogue-OTB (5.7%), which was consistent with our previous study (anti-OTA-based ELISA) [32]. No cross-reactivities were observed with other different mycotoxins, including AFB1, FB1, ZEN, PAT, CIT, and DON (<0.01%). These results demonstrated the good specificity of this MNPS-ELISA for the detection of OTA.

Optimization of the Electrochemical Biosensor Immunoassay
For the electrochemical test, several parameters were optimized to improve the performance. Different concentrations of H2O2 and hydroquinone (HQ) concentrations were assayed after the recognition event, where only tracer is added (0 OTA ppb). The use of 1mM of H2O2 and 1.5 μM of HQ resulted in an increased electrochemical signal. The optimum scanning frequency was −0.5 to −0.1 V and a scan speed of 100 mV s −1 .
The differential pulse voltammetry (DPV) scanning curve of benzoquinone (BQ) with different concentrations (seven units of twofold serial dilution from 0.5 μM) and the linear relationship between peak currents and concentrations are shown in Figure 3, which permitted the development of the electrochemical analytical method based on the catalytic conversion of HQ for the detection of OTA.

Specificity Study
The MNPs-ELISA based on MNPs-anti OTA and OTA-HRP had low cross-reactivity with the OTA analogue-OTB (5.7%), which was consistent with our previous study (anti-OTA-based ELISA) [32]. No cross-reactivities were observed with other different mycotoxins, including AFB 1 , FB 1 , ZEN, PAT, CIT, and DON (<0.01%). These results demonstrated the good specificity of this MNPS-ELISA for the detection of OTA.

Optimization of the Electrochemical Biosensor Immunoassay
For the electrochemical test, several parameters were optimized to improve the performance. Different concentrations of H 2 O 2 and hydroquinone (HQ) concentrations were assayed after the recognition event, where only tracer is added (0 OTA ppb). The use of 1mM of H 2 O 2 and 1.5 µM of HQ resulted in an increased electrochemical signal. The optimum scanning frequency was −0.5 to −0.1 V and a scan speed of 100 mV s −1 .
The differential pulse voltammetry (DPV) scanning curve of benzoquinone (BQ) with different concentrations (seven units of twofold serial dilution from 0.5 µM) and the linear relationship between peak currents and concentrations are shown in Figure 3, which permitted the development of the electrochemical analytical method based on the catalytic conversion of HQ for the detection of OTA.
recognition event, where only tracer is added (0 OTA ppb). The use of 1mM of H2O2 and 1.5 μM of HQ resulted in an increased electrochemical signal. The optimum scanning frequency was −0.5 to −0.1 V and a scan speed of 100 mV s −1 .
The differential pulse voltammetry (DPV) scanning curve of benzoquinone (BQ) with different concentrations (seven units of twofold serial dilution from 0.5 μM) and the linear relationship between peak currents and concentrations are shown in Figure 3, which permitted the development of the electrochemical analytical method based on the catalytic conversion of HQ for the detection of OTA.

Calibration Curve of the Electrochemical Biosensor Immunoassay
The calibration curve of the electrochemical biosensor immunoassay is shown in Figure 4. The equation of linearity is y = 0.3372x + 0.8324 with R 2 of 0.9805. The range of detection and the detection limit were 0.01-0.82 ng/mL and 0.007 ng/mL, respectively. In addition, 50% inhibition (IC50) was detectable at 0.10 ng/mL.

Calibration Curve of the Electrochemical Biosensor Immunoassay
The calibration curve of the electrochemical biosensor immunoassay is shown in Figure 4. The equation of linearity is y = 0.3372x + 0.8324 with R 2 of 0.9805. The range of detection and the detection limit were 0.01-0.82 ng/mL and 0.007 ng/mL, respectively. In addition, 50% inhibition (IC50) was detectable at 0.10 ng/mL. Previous studies have reported that in the detection of small-molecule analytes, a conjugation of analytes with HRP can be used for signal amplification [31]. Magnetic nanoparticle (MNPs), owing to their uniform diameter particles that can be homogenously distributed in colloidal suspension, reduced detection time and improved sensitivity. In this study, a MNPs-ELISA was developed using MNPs-anti OTA and OTA-HRP for signal amplification. As the electrochemical analysis has proven its ability in detecting traces of various analytes, a rapid and sensitive biosensor based on the MNPs-ELISA and screen-printed electrodes (MNPs-SPEs sensor) was established. By converting the signal generated by a chemical reaction to an electrical signal, the signal amplification was achieved once again. The detection limit (0.007 ng/mL) of the MNPs-SPEs sensor was significantly better than MNPs-ELISA (0.04 ng/mL), and the sensitivity increment was 10 old when compared with a common ELISA (0.07 ng/mL) established in our laboratory [32].
This presented MNPs-SPEs sensor offers higher sensitivity as compared with other methods in the detection of OTA, including fluorescence methods [33][34][35], enzyme immunoassay [30], fluorescence polarization (FP) analysis [36], and other different sensor methods [37][38][39][40][41]. Compared with other ultrasensitive techniques, such as fluorescence resonance energy transfer (FRET) aptasensor [42] and impedimetric aptasensor [43], our MNPs-SPEs sensor offers a similar detection limit, but in a shorter detection time. The comparative results of the performance parameters of these detection methods are shown in Table 3. The easy to use MNPs-SPEs sensor described herein can give precise and accurate results in a short time. This detection method is suitable not only for mycotoxins, but  Previous studies have reported that in the detection of small-molecule analytes, a conjugation of analytes with HRP can be used for signal amplification [31]. Magnetic nanoparticle (MNPs), owing to their uniform diameter particles that can be homogenously distributed in colloidal suspension, reduced detection time and improved sensitivity. In this study, a MNPs-ELISA was developed using MNPs-anti OTA and OTA-HRP for signal amplification. As the electrochemical analysis has proven its ability in detecting traces of various analytes, a rapid and sensitive biosensor based on the MNPs-ELISA and screen-printed electrodes (MNPs-SPEs sensor) was established. By converting the signal generated by a chemical reaction to an electrical signal, the signal amplification was achieved once again. The detection limit (0.007 ng/mL) of the MNPs-SPEs sensor was significantly better than MNPs-ELISA (0.04 ng/mL), and the sensitivity increment was 10 old when compared with a common ELISA (0.07 ng/mL) established in our laboratory [32].
This presented MNPs-SPEs sensor offers higher sensitivity as compared with other methods in the detection of OTA, including fluorescence methods [33][34][35], enzyme immunoassay [30], fluorescence polarization (FP) analysis [36], and other different sensor methods [37][38][39][40][41]. Compared with other ultrasensitive techniques, such as fluorescence resonance energy transfer (FRET) aptasensor [42] and impedimetric aptasensor [43], our MNPs-SPEs sensor offers a similar detection limit, but in a shorter detection time. The comparative results of the performance parameters of these detection methods are shown in Table 3. The easy to use MNPs-SPEs sensor described herein can give precise and accurate results in a short time. This detection method is suitable not only for mycotoxins, but also can be modified for other analytes, such as antibiotics, drugs, and pesticide residues.

Recovery Studies
Corn samples spiked with OTA at different levels (i.e., 1.25, 2.5, 5, 10, and 20 µg/kg) were analyzed in triplicate using the MNPs-SPEs sensor, and the results are shown in Table 4. The recovery rates were between 78.7% and 113.5%, and the relative standard deviations (RSDs) were 3.6-9.8%, which indicates that this MNPs-SPEs sensor assay was accurate and has good reproducibility.

Commercial Samples Analysis
A total of 56 dry commercial samples of corn, wheat, and feedstuff were analyzed in triplicate by the MNPs-SPEs sensor method and LC-MS/MS. The results of the positive samples are shown in Table 5. Using these 11 samples (5 corn, 4 wheat, and 2 feed) that were positive for ochratoxin, the relationship estimates between the MNPs-SPEs sensor and LC-MS/MS methods for OTA detection was assessed by regression analysis, MNPs-SPEs sensor = 1.7539 + 0.9146LC-MS/MS (R 2 = 0.8149). It revealed that there is a good agreement between these two methods ( Figure 5). This also demonstrate that our MNPs-SPEs sensor system is applicable for the detection of OTA.
by the MNPs-SPEs sensor method and LC-MS/MS. The results of the positive samples are shown in Table 5. Using these 11 samples (5 corn, 4 wheat, and 2 feed) that were positive for ochratoxin, the relationship estimates between the MNPs-SPEs sensor and LC-MS/MS methods for OTA detection was assessed by regression analysis, MNPs-SPEs sensor = 1.7539 + 0.9146LC-MS/MS (R 2 = 0.8149). It revealed that there is a good agreement between these two methods ( Figure 5). This also demonstrate that our MNPs-SPEs sensor system is applicable for the detection of OTA.

Conclusions
In this work, a screen-printed electrodes (SPEs)-based electrochemical biosensor (MNPs-SPEs sensor) was successfully engineered for the detection of ochratoxin A in cereal and feed samples utilizing magnet nanoparticles (MNPs) and a small molecule-HRP conjugate (OTA-HRP) for signal enhancement. For MNPs-SPEs sensor systems, the detection limit was 0.007 ng/mL, the detection range was 0.01-0.82 ng/mL, and the IC50 was 0.10 ng/mL. The recovery rates in spiked cereal samples spread between 78.7-113.5%, and the RSD values were all <15%. Analysis of commercial samples using this MNPs-SPEs sensor and LC-MS/MS revealed a good correlation between these two methods. The lower limit of detection and shorter reaction time of less than one hour makes this new assay an excellent alternative to existing conventional methods for the detection of trace amounts of OTA in agroproducts. Additionally, this platform can also be adapted for the detection of other small targets, as well as offering wide applications in food safety-related fields.

Equipment
The equipment used in this study were the following: the 37 • C incubator from Thermo Scientific (Waltham, MA, USA); the horizontal shaker (Vortex 4 basic) from IKA (Staufen, Germany); the magnetic separator (MS-12) from Bangs laboratories (Fishers, IN, USA); and the Spectra Max M 2 micro-plate reader from Molecular Devices (Sunnyvale, CA, USA). Electrochemical measurements were performed with a PC-controlled CHI-832 electrochemical analyzer (Chenhua, Shanghai, China).

Synthesis of the OTA and Antibody Conjugates
OTA-HRP was prepared by a two-step approach with slight modifications [44]. OTA (2.0 mg) was dissolved in 350 mL of anhydrous tetrahydrofuran (THF) and then 3.0 mg of N-hydroxy-succinimide (NHS) and 12.0 mg of N,N -dicyclohexylcarbodiimide (DCC) were added, followed by gentle shaking at room temperature (RT) for 12 h. The reaction mixture was centrifuged (10,000× g, 15 min). The supernatant was dried and dissolved in 0.3 mL dimethyl sulfoxide (DMSO).
HRP (2 mg) was dissolved in 2.0 mL of 0.13 M phosphate buffer (PBS, pH 8.0), and the activated OTA (0.3 mL) was added drop-wise. The reaction was proceeded by vigorous shaking at RT for 6 h and then dialyzed extensively against 0.01 M/L of PBS (pH 7.4) at 4 • C. The OTA-BSA conjugate was synthesized as previously reported [32]. In brief, OTA was dissolved in THF and NHS and DCC were added, followed by gentle shaking at room temperature (RT). The reaction mixture was centrifuged, the supernatant was dried, and the residue was dissolved in DMSO. BSA was dissolved in 0.13 M phosphate buffer (pH 8.0). The activated OTA was added drop-wise to the BSA solution. The reaction was allowed to proceed by vigorous shaking at RT and then dialyzed extensively against 0.01 M/L of phosphate buffered saline (PBS, pH 7.4). OTA-BSA-HRP and Anti OTA-HRP were synthetized as recommended by the supplier (EL0156, Huzhou InnoReagents, Huzhou, China).

Preparation of Magnetic Nanoparticles Conjugates
The immunomagnetic beads (MNPs-Anti OTA) were synthetized and the coating efficiency was measured as described in our previous study [22]. OTA-BSA-coated magnetic nanoparticles (MNPs-BSA-OTA) were prepared in similar fashion as the immunomagnetic beads. In brief, OTA-BSA (100 µg) in coating buffer (100 µL) was added to the activated magnetic nanoparticles and incubated for 2 h at 4 • C with slow tilting rotation. After quenching, the non-reacted activated carboxylic acid groups with quenching buffer, the coated nanoparticles were resuspended in the storage buffer for later use. The coating/quenching/storage buffers were the same as those mentioned in the preparation of MNPs-Anti OTA.

Development of MNP-ELISAs
Three different types of MNPs-ELISA were developed using the magnetic nanoparticles to assay their effects on signal release and performance. Figure 1 shows an illustration comparing these MNPs-ELISA variants.
As shown in Figure 1A, MNPs-BSA-OTA diluted with the storage buffer (10 µL), 70 µL of the Anti OTA-HRP solution, and 70 µL of the diluted test sample extracts (or OTA standard solution) were added to the 96-well plate in triplicate wells. Then, shaking the plate at 1000 rpm for 45 min at 37 • C was followed by placing the samples on a magnetic base to segregate the nanoparticles. The immunomagnetic nanoparticle complexes thus separated were washed thrice. The substrate 3,3 ,5,5 -tetramethylbenzidine (TMB,100 µL) was added. Stop solution (2 M H 2 SO 4 , 50 µL) was pipetted after 10 min of rotatory incubation at 37 • C, and OD450 was read on a Spectra Max M2 micro-plate reader. The calibration curve for MNPs-bsELISA was prepared in GraphPad 5 software with the x-axis representing the log concentration of OTA (ng/mL) and the y-axis, the inhibition rate. The inhibition rate is calculated as one minus the ratio of the OD450 of standard OTA solutions in PBS to the OD450 of PBS (0.01 M, pH 7.4). Figure 1B, compared with Figure 1A, the variant involved replacing MNPs-BSA-OTA with MNPs-Anti OTA. For the subsequent steps, diluted sample extracts or standard OTA solutions were mixed with an equal volume of OTA-BSA-HRP (70 µL), and then, the mixtures (150 µL) were transferred to the 96-well plate. Other steps were similar to those described in Figure 1A.

As shown in
In Figure 1C, as compared with Figure 1B, OTA-BSA-HRP was replaced with OTA-HRP. For downstream assay, diluted sample extracts or standard OTA solutions were mixed with an equal volume of OTA-HRP (70 µL), and then, the mixtures (150 µL) were transferred to the 96-well plate, with other steps similar to those described in Figure 1A.

Optimization of MNPs-ELISAs
Performance enhancement of these three types of MNPs-ELISA was obtained by optimizing the dilutions of MNPs-Anti OTA, MNPs-BSA-OTA, concentrations of OTA-BSA-HRP, Anti OTA-HRP, and OTA-HRP, and times for the competition reaction were as described in our previous study [22]. The concentrations/dilution ratios of immune-reagents for different MNPs-ELISAs were optimized by checkerboard titration design with an OD450 value of about 1.0. Incubation times for the competition reaction were 30, 45, 60, and 90 min.

Specificity Study
To evaluate the specificity of MNPs-ELISA, cross-reactivities (CRs) with seven other different mycotoxins (OTB, AFB 1 , FB 1 , ZEN, DON, PAT, and CIT) were determined, as reported previously [22]. In brief, using the developed MNPs-bsELISA, the calibration curve of OTA with different concentrations was established first. Then, different concentrations of each analyte (instead of OTA) were used as potential binding competitors, and the calibration curve including the IC50 (50% inhibition) for each analyte was calculated respectively. Cross-reactivity was calculated as percent inhibition using the following formula: IC50 of OTA/IC50 of other mycotoxins × 100% [32].

Development of the Electrochemical Biosensor Immunoassay
Electrochemical biosensor immunoassay is developed based on the MNPs-ELISA. After the competition reaction of the MNPs-ELISA is complete, immunomagnetic nanoparticle complexes were resuspended in 100 µL of PBS buffer and transferred onto the surface of a screen-printed electrode on which nanoparticles were immobilized by placing a magnet at the bottom. The immobilized complexes were then rinsed thoroughly with ultrapure water and dried with nitrogen. After that, 100 µL of a solution containing the enzymatic substrate (1 mM H 2 O 2 ) and the electrochemical mediator (1.5 µM HQ) in PBS buffer (50 mM) was deposited on the electrode surface with volume enough for covering the three-electrode system. The current response was measured using differential pulse voltammetry (DPV). All DPV measurements were performed in the potential range of −0.5 to −0.1 V and a scan speed of 100 mV s −1 . The extent of affinity reaction was evaluated by the addition of an electrochemical mediator whose reduction on the electrode surface is directly related to the activity of the enzyme tracer. Figure 1 shows the schematic illustration of the magnetic nanoparticles-based electrochemical biosensor (MNPs-SPEs sensor) test procedure.

Recovery Study and Comparison of Detection in Commercial Samples by LC-MS/MS
OTA-free corn samples (tested by LC-MS/MS) were ground and dried by overnight incubation in a 60 • C incubator and spiked with a standardized solution of OTA at different concentrations. Then, the spiked samples were vortexed for 10 min and incubated at RT overnight. Next, 25 mL of methanol/water (70:30, v/v) were added to each sample (5 g) and vortexed for 10 min. The samples were centrifuged at 3000g for 10 min, and the supernatants were diluted seven-fold with PBS to minimize the influence of the solvents. Each sample was analyzed in triplicate.
Dry commercial samples (including corn, wheat, and feedstuff) were analyzed by the developed MNPs-SPEs sensor and LC-MS/MS in parallel. Each sample was tested in triplicate to calculate the standard deviation. For the detection by the electrochemical immunoassay, the samples were extracted as the spiked samples. Validated procedures for LC-MS/MS were adopted as those described previously [45].
The correlation between these two methods was calculated using linear regression (Microsoft Excel software, Redmond, WA, USA; 2016 version).
Author Contributions: X.Z., Z.W., R.S., and W.F. conceived of and designed the experiments; X.Z., T.C., and H.X. conducted the experiments and analyzed the data; H.S. and W.F. contributed the materials; X.Z. wrote the manuscript; and N.P. and H.S. supervised the work and revised the manuscript.